mc = 10
#--------------------------------------------------#
#training
#data2: combine 2014, first 80% for train, 20% for test
data2013 = list(DictReader(open("pbp-2013.csv", 'r')))
data2014 = list(DictReader(open("pbp-2014.csv", 'r')))
data2015 = list(DictReader(open("pbp-2015.csv", 'r')))
dataList = [data2013, data2014, data2015]
dataName = ["2013","2014","2015"]

o = DictWriter(open("rfClassifier-mc.csv", 'w'), ["dataName", "classifier", "percent", "score", "OmniScore", "Type1-A/A/Good","Type2-A/B/Bad",  "Type3-A/B/Good", "Type4-A/A/Bad"])
o.writeheader()

#---------------------------------#
for dataindex in range(len(dataList)):
    pbp2014 = NewPbpExtractor()
    pbp2014.buildFormationList(dataList[dataindex])
    feature, target = pbp2014.extract4Classifier(dataList[dataindex])

    dataLength = feature.shape[0]
    dataLength80 = round(dataLength * 0.8)
    X_train = feature[0:dataLength80,:]
    X_test = feature[(dataLength80+1):dataLength,:]
    y_train = target[0:dataLength80]
    y_test = target[(dataLength80+1):dataLength]


    class2014 = classifierEvaluation()

    Bscore, Bnum, BtypeNum, BomniScore = class2014.Score(y_test, y_test)
    baseline = {'dataName': dataName[dataindex] ,'classifier': 'baseline', 'percent': Bscore/float(BomniScore),'score': Bscore,'OmniScore': BomniScore, 'Type1-A/A/Good': BtypeNum[0], 'Type2-A/B/Bad': BtypeNum[1], 'Type3-A/B/Good': BtypeNum[2], 'Type4-A/A/Bad': BtypeNum[3]}
from sklearn.cross_validation import train_test_split


#--------------------------------------------------#
#training
#data2: combine 2014, first 80% for train, 20% for test
data2013 = list(DictReader(open("pbp-2013.csv", 'r')))
data2014 = list(DictReader(open("pbp-2014.csv", 'r')))
data2015 = list(DictReader(open("pbp-2015.csv", 'r')))
dataList = [data2013, data2014, data2015]
dataName = ["2013","2014","2015"]


#train data 2015
data13and14 = data2013+data2014
pbp2014 = NewPbpExtractor()
pbp2014.buildFormationList(data13and14)
feature, target = pbp2014.extract4Classifier(data13and14)
X_train = feature
y_train = target

#test data 2015
pbp2014 = NewPbpExtractor()
pbp2014.buildFormationList(data2015)
feature, target = pbp2014.extract4Classifier(data2015)
X_test = feature
y_test = target

o = DictWriter(open("Bayes.csv", 'w'), ["dataName", "classifier", "percent", "score", "OmniScore", "Type1-A/A/Good","Type2-A/B/Bad",  "Type3-A/B/Good", "Type4-A/A/Bad"])
o.writeheader()